Dynamic Optimization of Individual Financial Well-Being

Sammanfattning

This thesis develops a dynamic programming framework for individualized financialadvisory in a Swedish banking context. The study focuses on young adults aged 18–30 who face several competing financial goals, including maintaining an emergencybuffer, saving for a housing down payment, building short-term savings, andaccumulating long-term savings. The purpose is to examine how financial well-being can be represented through measurable financial goals and used in a dynamicoptimization model. The aim is to study whether the model produces financiallyinterpretable allocation behavior under exogenous risks.Financial well-being is represented through measurable components. Thesecomponents are aggregated into alternative utility specifications and embedded ina finite-horizon dynamic programming model. The model incorporates stochasticelements such as unemployment, financial crises, income variation, and short-termspending. Parameter values are calibrated using Swedish public statistics, officialinstitutional sources, return assumptions from Handelsbanken, and relevant literature.The model is solved numerically in Python, and the resulting policy allocations areevaluated using Monte Carlo simulations and sensitivity analysis.The results suggest that the model can represent trade-offs between liquidity,goal attainment, and long-term financial development. Specifications with a strongerweight on housing savings reach the down payment goal more frequently and faster,while more balanced specifications perform better when income is constrained andfinancial resilience becomes more important. The inclusion of a housing-goal bonusfurther strengthens goal-oriented behavior and improves down-payment outcomes.However, the results are sensitive to parameter choices, particularly the utility weightsand the housing bonus.The thesis illustrates how dynamic programming can provide a structured andtransparent framework for analyzing financial allocation decisions across severalcompeting goals. While the model relies on simplifying assumptions and a limiteddefinition of financial well-being, it illustrates how optimization-based methods couldserve as a conceptual basis for more individualized advisory tools in banking practice.